Abstract:
7-Methylxanthine, a derivative of caffeine (1,3,7-trimethylxanthine), is
a high-value compound that has multiple medical applications,
particularly with respect to eye health. Here, we demonstrate the
biocatalytic production of 7-methylxanthine from caffeine usingEscherichia coli strain MBM019, which was constructed for
production of paraxanthine (1,7-dimethylxanthine). The mutantN -demethylase NdmA4, which was previously shown to catalyzeN 3-demethylation of caffeine to produce
paraxanthine, also retains N 1-demethylation
activity toward paraxanthine. This work demonstrates that whole cell
biocatalysts containing NdmA4 are more active toward paraxanthine than
caffeine. We used four serial resting cell assays, with spent cells
exchanged for fresh cells between each round, to produce 2,120 μM
7-methylxanthine and 552 μM paraxanthine from 4,331 μM caffeine. The
purified 7-methylxanthine and paraxanthine were then isolated viapreparatory-scale HPLC, resulting in 177.3 mg 7-methylxanthine and 48.1
mg paraxanthine at high purity. This is the first reported strain
genetically optimized for the biosynthetic production of
7-methylxanthine from caffeine.
Keywords: 7-Methylxanthine, Caffeine, Biocatalysis,N -demethylase
7-Methylxanthine, a purine alkaloid, is a derivative of the well-known
compound caffeine (1,3,7-trimethylxanthine). Like caffeine,
7-methylxanthine is an adenosine receptor antagonist and is capable of
crossing the blood-brain barrier (Hung et al., 2018). These features
make 7-methylxanthine and other methylxanthines attractive as scaffolds
for the design and synthesis of more complex compounds (Nivedita Singh,
2018), like N-heterocyclic carbenes (Valdés et al., 2018; Zhang et al.,
2015), that can be used for applications such as fine-tuning the potency
and specificity of receptor antagonism for use in cancer and
neurodegenerative therapies (Malki et al., 2006; Rogozin et al., 2006).
While many of the specific health benefits commonly assigned to caffeine
could arguably belong instead to one or more of its derivatives, the
health benefits of 7-methylxanthine are much more clearly defined. Most
notably, 7-methylxanthine has been shown to treat and prevent myopia
progression (Cui et al., 2011; Nie et al., 2012). While the exact
mechanisms of action have not yet been fully characterized, oral
application of 7-methylxanthine has been observed to strengthen the
sclera (Cui et al., 2011), reducing the potential for axial elongation
and even showing a moderate decline in the rate of axial elongation in
myopic children (Trier et al., 2008). Additionally, methylxanthine
derivatives, such as 1,3-dipropyl-7-methylxanthine, have been found to
increase the sensitivity of lung carcinoma cell lines through
modification to cell cycle checkpoints and by inducing apoptotic
responses (Malki et al., 2006). Toxicity studies of 7-methylxanthine
have concluded that the compound is nontoxic. Even at concentrations as
high as 2,000 mg/kg body weight, 7-methylxanthine induces no observable
changes in behavior and is considered safe for long-term, chronic use
(Singh et al., 2019; Singh et al., 2020).
7-Methylxanthine is not found in nature as frequently as other
methylxanthines, such as caffeine, theobromine (3,7-dimethylxanthine),
and theophylline (1,3-dimethylxanthine), but it can still be found in
low concentrations in plants as an intermediate during the synthesis of
caffeine (Maureen McKeague, 2016). 7-Methylxanthine has also been
generated in Escherichia coli through the N -demethylation
of theobromine using N -demethylase genes isolated and
characterized from Pseudomonas putida CBB5 (K. H. R. Algharrawi
& Subramanian, 2020). Other studies have reported that varying
combinations of the ndmABCDE genes discovered in P. putidaCBB5 could be used to produce 3-methylxanthine from theophylline (K. H.
Algharrawi et al., 2015) and theobromine from caffeine (K. H. Algharrawi
et al., 2017).
We have recently constructed E. coli strain MBM019 toN- demethylate caffeine primarily to paraxanthine
(1,7-dimethylxanthine), a caffeine metabolic intermediate (Mock et al.,
In Press 2022) with the potential to treat and prevent Parkinson’s
Disease (Janitschke et al., 2021; Victorino et al., 2021). This
engineered bacterium harbors a mutant N -demethylase gene,ndmA4 (Kim et al., 2019; Mills et al., 2021), originally designed
for the sole purpose of shifting the primary N -demethylation
product from theobromine to paraxanthine (Mills et al., 2021). Strain
MBM019 also overexpresses the ndmDP1 gene as well as two
formaldehyde-degrading genes, frmA and frmB (Fig S1).
NdmDP1 is a truncated version of NdmD, a reductase required forN -demethylation through the transfer of electrons from NADH (R.
Summers et al., 2014; R. M. Summers et al., 2012), that provides a
higher N -demethylation activity than the wild-type NdmD (Mock et
al., In Press 2022). The FrmAB enzymes are native to E. coli and
catalyze the NAD+-dependent degradation of
formaldehyde, a by-product of N -demethylation
(Mock et al., In Press 2022), thus
generating an NADH/NAD+ recycle system within the
cell. In our study to produce paraxanthine from caffeine, we also
observed the generation of a small amount of 7-methylxanthine (Mock et
al., In Press 2022). Here, we describe an alternate pathway to
7-methylxanthine via paraxanthine by nearly complete conversion of
caffeine using strain MBM019 (Fig 1), establishing a biosynthetic
process for consuming an environmental contaminant and sequentially
producing two high-value compounds.
Production of 7-methylxanthine from caffeine by strain MBM019 was
optimized in a 15 mL resting cell assay reaction in which the reaction
supernatant was recycled three times with fresh cells for a total of
four rounds of reaction (Fig 2, Table S1 & Fig S2). Each reaction was
carried out at the previously-optimized conditions of a cell
OD600 of 50 and initial caffeine concentration of 5 mM
(Mock et al., In Press 2022). After the first round of reaction, 1,686 ±
121 μM caffeine was consumed, resulting in 906 ± 26 μM paraxanthine and
350 ± 19 μM 7-methylxanthine (Table S1). The purpose of multiple
reactions using fresh cells with reused supernatant stems from the
observation that the reaction slowed greatly after five hours and
ultimately plateaued with 3,344 ± 29 μM caffeine remaining (Fig 2 Round
1 & Table S1). We then hypothesized that the addition of fresh cells
would further increase conversion of caffeine to paraxanthine and
7-methylxanthine. Thus, the reaction supernatant was recycled with fresh
cells (OD600 = 50), constituting the second reaction and
resulting in consumption of an additional 1,862 ± 20 μM caffeine. The
concentration of paraxanthine increased slightly to 1,033 ± 7 μM, while
the concentration of 7-methylxanthine increased to 1,426 ± 36 μM. Both
caffeine and paraxanthine decreased in further rounds of recycled
reactions as 7-methylxanthine continued to increase (Fig 2 & Table S1).
After four cycles, the reaction mixture contained 231 ± 10 μM caffeine,
274 ± 4 μM paraxanthine, and 2,614 ± 21 μM 7-methylxanthine (Fig 2 &
Table S1).
After process optimization was complete, the reaction was scaled-up for
purification purposes. Strain MBM019 was grown in 4 L LB media for use
in one cycle of a 640 mL reaction with the optimized conditions
determined from the 15 mL reactions. Spent cells were removed and
freshly grown cells were added to a final OD600 of 50
between each of the four cycles, resulting in a final supernatant volume
of 580 mL after removing cells from the fourth cycle. After four
large-scale reaction cycles, caffeine was degraded to a final
concentration of 669 μM with concomitant production of 552 μM
paraxanthine and 2,120 μM 7-methylxanthine. Overall, 86.6 mol% of
caffeine was consumed. Conversion to 7-methylxanthine accounted for 42.6
mol% of caffeine and conversion to paraxanthine accounted for 11.0
mol%, representing a total of 53.6 mol% of the converted caffeine. The
remaining 33.0 mol% of unaccounted product may have been converted to
other compounds, such as 1-methylxanthine. This theory is supported by
the presence of several unconfirmed peaks seen in the HPLC chromatograph
(Fig S2) (Mock et al., In Press 2022), as well as observed enzyme
promiscuity that has been previously characterized in Ndm enzymes when
reacted in vivo (Mock et al., 2021).
Products were isolated from the
reaction supernatant using two rounds of preparatory-scale HPLC (Fig
S3). HPLC purification resulted in
separation efficiencies of 92.73% for 7-methylxanthine and 98.73% for
paraxanthine (Table S2). Following purification, the compounds were
dried to a powder and collected, resulting in the recovery of 177.3 mg
7-methylxanthine and 48.1 mg of paraxanthine (Table S2 & Fig S4).
7-Methylxanthine and paraxanthine purity was analyzed by HPLC using
authentic standards and the retention times were confirmed to be the
same (Fig S4). 1H-NMR was also used to confirm the
identity of the biologically produced 7-methylxanthine and paraxanthine
(Fig S5). The presence of proton peaks correlating with 7-methylxanthine
were confirmed at δ 11.43 (1H) and δ 10.86 (1H) corresponding to –NH, δ
7.88 (1H) corresponding to -C=CH, δ 3.82 (3H) corresponding to the
–CH3 group. The presence of proton peaks correlating
with paraxanthine were confirmed at δ 11.83 (1H) corresponding to –NH,
δ 7.92 (1H) corresponding to -C=CH, δ 3.86 (3H) and δ 3.18 (3H)
corresponding to both –CH3 groups. Peaks δ 3.32 and δ
2.50 in both chromatograms are water and DMSO, respectively. There is a
very small additional peak observed in the 7-methylxanthine
chromatograph just below δ 2 that is believed to be residual acetic
acid.
During the four-cycle production reaction, we observed a maximum
paraxanthine concentration of approximately 1 mM, after which
7-methylxanthine concentration began to increase rapidly (Fig. 2). This
plateau in paraxanthine concentration was surprising, as the NdmA4
mutant was designed for N 3-demethylation of
caffeine to paraxanthine, and a single round of reaction yielded
paraxanthine as the major product. Data from the multi-cycle reaction
has led us to hypothesize that NdmA4 is more active toward paraxanthine
than caffeine. Indeed, a small-scale MBM019 resting cell assay converted
516.6 ± 49.6 μM paraxanthine to 459.1 ± 27.2 μM 7-methylxanthine over
five hours, compared with conversion of 345.7 ± 19.3 μM caffeine to
204.9 ± 6.0 μM paraxanthine and 73.4 ± 7.6 μM 7-methylxanthine over the
same time period (Fig. 3). Use of baffled flasks to improve oxygenation
in the resting cell assay yielded no improvement (data not shown),
indicating that oxygen is not limiting. The wild-type NdmA is anN 1-demethylase capable of fully converting 1 mM
caffeine to theobromine in 90 minutes, and NdmA4 has retained thatN 1-demethylation activity toward paraxanthine,
albeit at a lower rate, resulting in a single enzyme capable of
producing 7-methylxanthine from caffeine. Further enzyme optimization
via mutagenesis and the subsequent study of enzyme kinetics will be
required to generate an enzyme with improved activity toward caffeine
and yield of 7-methylxanthine.
In summary, we have demonstrated the ability to produce 7-methylxanthine
from caffeine through four serial resting cell reactions using the
previously optimized E. coli MBM019 strain, followed by HPLC
purification of the paraxanthine (minor product) and 7-methylxanthine
(major product) generated. The identity of these products was further
confirmed by 1H-NMR. This study also demonstrated that
cells expressing the NdmA4 mutant enzyme are more active toward
paraxanthine than caffeine. To our knowledge, this is the first report
of the biocatalytic production of 7-methylxanthine from caffeine.
Cell Growth, Protein Expression and Resting Cell Assays
The E. coli strain MBM019 was grown in LB medium, and gene
expression was induced as previously described by Mock et al .
(Mock et al., In Press 2022; Mock et al., 2021). For more details,
please reference the Supplemental Material. The cells were then
harvested by centrifugation at 10,000 x g for 10 min at 4℃ and
washed twice in ice cold 50 mM potassium phosphate (KPi)
prior to assay. For reactions designated for product isolation,
four 2.8 L Fernbach flasks each
containing 1 L LB medium was used for cell growth. After washing, cells
were resuspended in 10 mL of ice cold 50 mM KPi.
Reactions for supernatant recycling optimization with fresh cells were
carried out in triplicate at an OD600 of 50, starting
substrate concentration of 5 mM, and at a volume of 15 mL. Subsequent
reaction substrate concentrations were dependent on the extent of the
previous reaction. Reactions for the comparison of caffeine and
paraxanthine as substrates were carried out in triplicate at an
OD600 of 5, substrate concentration of 1 mM, and at a
volume of 2 mL. A single large-scale reaction for purification was
conducted at the maximum volume possible while still retaining the
required OD600 of 50 for large-scale reactions with a
caffeine concentration of 5 mM.
HPLC Separation
A detailed description of separation methods is provided in the
Supplemental Material. Briefly, the harvested supernatant was filtered
through a 0.2 μm filter, and methanol (MeOH) was added to the
supernatant to reduce changes in MeOH concentration within the system
during purification. The first round of HPLC purification was designated
for the removal of caffeine and for the crude separation and collection
of paraxanthine and 7-methylxanthine using a 15% MeOH mobile phase (Fig
S3). The collected solutions were concentrated via rotary evaporation at
70℃ and 200-220 mbar. A second round of HPLC purification with 5% MeOH
as mobile phase was required for complete purification of both the
7-methylxanthine and the paraxanthine solutions. Both purified solutions
were concentrated by rotary evaporation and then dried to a powder (Fig
S4 & S5).
Acknowledgments: The authors thank Dr. Ken Belmore and the
University of Alabama Department of Chemistry and Biochemistry for
assistance with the NMR. This work was supported by University of
Alabama research funds. M.B. Mock is supported by the U.S. Department of
Education as a GAANN Fellow (P200A180056).
Conflict of interests: The authors declare no conflicts of
interest.
Data availability statement: The authors confirm that the data
supporting the findings of this study are available within the article
and the supporting information.